News
Similarly, on HumanEval to evaluate Python code generation and CruxEval to test Python output prediction, the model bested the competition with scores of 81.1% and 51.3%, respectively.
Results that may be inaccessible to you are currently showing.
Hide inaccessible results